1. Pituitary somatotrophs secrete growth hormone spontaneously. This is controlled by burstign oscillations in membrane potential (that is, action potentials occur in bursts separated by silent periods). This grouping together of action potentials produces an oscillatory but generally elevated calcium level, which is more efficient for secretion than isolated action potentials. Modulation up or down from this basic level is mediated by hypothalamic input of growth hormone releasing hormone (GHRH) and somatostatin (SRIF), respectively. We carried out mathematical modeling in parallel with experiments in the laboratory of S. Stojilkovic (NICHD) using channel blockers and ion substition experiments to elucidate both the oscillation mechanism and its modulation. The key element of the model is negative feedback on large conductance (BK) calcium-activated potassium channels. We found that up-regulation could be explained by activation of an inward, non-selective cation current and/or deactivation of an inward rectifier potassium (KIR) channel. The reverse is true for down regulation. Finally, we found that the natural variability of oscillation period (a few seconds to tens of seconds) could be accounted for by different levels of BK conductance. The identity of the inward cation current is still unknown, but some of its properties have been clarified, such as that it is blocked by magnesium and carries sodium, which may help in ultimately identifying it. See Ref. # 3.[unreadable] [unreadable] 2. Like the pituitary somatotrophs described above, other pituitary cells (lactotrophs, which secrete prolactin; and corticotrophs, which secrete adrenocorticotropic hormone) show bursting oscillations that are similar in appearance. For example, the spikes on the plateau are small and noisy. We were interested to contrast models for such oscillations with models for more classically studied bursting in insulin-secreting pancreatic beta-cells. The latter look somewhat different, with much larger spikes and often, but not always, longer lasting plateaus. However, such quantitative differences are subtle and not reliable for distinguishing fundamental mechanisms. Instead, we found that the two types of oscillations